Dominique Bagnard

Plexins: making links to the cytoskeleton and reducing the GAPs in our knowledge

During development of the nervous system, the establishment of axonal projections requires tight control of the extraordinarily complex machinery of growth cones — the specialized axonal structures enabling pathway selection. The discovery of several families of guidance molecules displaying attractive and repulsive properties provided new insights into the molecular nature of axonal guidance. Among these guidance factors, the functions of different members of the semaphorin family have been well characterized. In particular, semaphorin 3A (Sema3A) is a potent inducer of growth cone collapse by a mechanism that results in the depolymerization of F-actin. Several studies have demonstrated the involvement of a receptor complex for the transduction of the Sema3A signal. It appears that the first-reported receptor, neuropilin-1, acts as a ligand-binding subunit requiring additional receptor components such as members of the plexin family for signal transduction.Although understanding of the ligand–receptor dialogue seems to be well advanced, little is known about the link between the receptors and the structural changes of the cytoskeleton during chemorepulsion. Now, two groups have demonstrated that plexins are able to modulate the activity of small GTPases that are key regulators of the actin cytoskeleton. Based on sequence homology between the intracellular domain of plexin and Ras GTPase-activating proteins (GAPs), Rohm et al.1xThe semaphorin 3A receptor may directly regulate the activity of small GTPases. Rohm, B et al. FEBS Lett. 2000; 486: 68–72Abstract | Full Text | Full Text PDF | PubMed | Scopus (126)See all References1 discovered that mutating two conserved arginine residues was sufficient to block Sema3A signal transduction through plexin-A1. They also provided evidence that the intracellular domain of plexins binds to specific Rho-like GTPases, allowing a potential activation of multiple signaling cascades. In the same way, Vikis et al.2xThe semaphorin receptor plexin-B1 specifically interacts with active Rac in a ligand-dependent manner. Vikis, H.G et al. Proc. Natl. Acad. Sci. USA. 2000; 97: 12457–12462Crossref | PubMed | Scopus (138)See all References2 have demonstrated that the cytosolic domain of plexin-B1 interacts directly with activated Rac.These first demonstrations of a GAP activity related to a transmembrane protein that is essential for the response to axonal guidance molecules open new perspectives in the understanding of the relationship between the detection of environmental cues and making appropriate cytoskeletal modifications.

Membrane traffic: Role in polarity and locomotion

In many ways, the neuromuscular junction has served as the grand paradigm for the architecture of a synapse. The particular morphological arrangements of pre- and postsynaptic elements at this junction, in particular the striking regular invaginations of the postsynaptic membrane, termed T-tubules, have been well described. Despite extensive ultrastructural studies, the molecular mechanisms involved in T-tubule formation are not yet fully understood. Astonishing insights by de Camilli and colleagues now reveal a novel and unexpected role for the lipid-binding protein amphiphysin in T-tubule biogenesis in muscle [9xAmphiphysin 2 (BIN1) and T-tubule biogenesis in muscle. Lee, E. et al. Science. 2002; 297: 1193–1196Crossref | PubMed | Scopus (216)See all References[9].In striated skeletal muscle, excitation–secretion coupling at the postsynaptic membrane requires inwardly extending plasma membrane folds, the T-tubules, which relay the electrical signal to the sarcoplasmic reticulum, the major Ca2+ storage site. The resulting increase in intracellular Ca2+ then initiates myofibril contraction. It therefore seems clear that development of functional muscle cells crucially depends on T-tubule biogenesis. de Camilli and colleagues have now tackled the question of how T-tubules are formed.Based on the reported observation that a splice variant of the lipid-binding protein amphiphysin 2, a protein previously implicated in clathrin- and dynamin-dependent endocytosis of presynaptic membranes, is highly concentrated in adult striated muscle, the authors set out to investigate its function. Strikingly, when overexpressed in non-muscle cells, GFP-tagged muscle amphiphysin 2 (M-Amph 2) is sufficient to trigger the formation of numerous narrow tubules that emanate from the plasma membrane. Tubulating activity in these cells is dependent on the lipid-associating BAR domain of M-Amph 2 and on the presence of a muscle splice-variant-specific exon, encoding a module binding to phosphatidylinositol (4,5)-bisphosphate (PIP2). Consistent with a proposed role of caveolin 3 in myogenesis, this protein becomes co-recruited along with dynamin 2 into M-Amph-2-covered, PIP2-rich tubules. These finding are matched by the concomitant upregulation of M-Amph 2 expression and PIP2 formation during muscle differentiation in cultured cells in vitro. Suppression of M-Amph 2 biosynthesis by RNA interference abrogates T-tubule formation and muscle cell differentiation. Collectively, these elegant studies demonstrate that expression of a single protein, M-Amph 2, is required and sufficient to induce formation of T-tubules and imply a physiological function for the previously observed membrane-tubulating activity of amphiphysin seen in vitro.The new data not only provide the first mechanistic and structural insights into T-tubule biogenesis but also suggest a common mechanism for how the membrane deforming properties of amphiphysin might contribute to diverse cellular processes, ranging from clathrin-mediated endocytosis at synapses to formation of T-tubules in striated muscle. Alternative splicing appears to direct amphiphysin to the correct membrane site, enabling its BAR domain to exert a tubulating activity, functioning either in the generation of dynamin-coated tubular endocytic necks or of T-tubules. Consistent with the new study, a Drosophila mutant lacking amphiphysin displays defects in T-tubule biogenesis. Hopefully, these new insights will pave the way for a more thorough understanding of muscle formation and the treatment of devastating diseases such as muscular dystrophy.

Netrin and the GTPases

Guidance molecules such as the netrins, the semaphorins or ephrins are key regulators of development of the central nervous system. These signals are thought to modulate neurite extension by providing instructive cues that act as attractants or repellents in different cell types. Current research focuses on the identification of the intracellular pathway leading to the regulation of the actin cytoskeleton required for such adapted cell behaviors. The small GTPases of the Rho family (RhoA, Rac1 and Cdc42) are largely involved in the transduction cascade ensuring the formation of filopodia (Cdc42) and lamellipodia (Rac1) of elongating growth cones. Therefore, it was logical to investigate recruitment of small GTPases during axonal guidance in response to semaphorins or ephrins. In this way, the interactions between Netrin-1 (a bifunctional guidance signal), DCC (‘deleted in colorectal cancer’, the receptor for Netrin-1) and Rho family members Rac1 and Cdc42 have recently been explored [1.xThe Netrin-1 receptor DCC promotes filopodia formation and cell spreading by activating Cdc42 and Rac1. Sherakabi, M. and Kennedy, T.E. Mol. Cell. Neurosci. 2002; 19: 1–17PubMedSee all References, 2.xSee all References].Sherakabi and Kennedy [1xThe Netrin-1 receptor DCC promotes filopodia formation and cell spreading by activating Cdc42 and Rac1. Sherakabi, M. and Kennedy, T.E. Mol. Cell. Neurosci. 2002; 19: 1–17PubMedSee all References][1] analyzed the consequence of inactivation of GTPases in different cell lines and growth cones following netrin-1-dependent filopodia extension. In their model, induced expression of DCC triggered an increase in the number of filopodia and cell-surface area when cells were exposed to netrin-1. Coexpression of DCC with dominant-negative forms of Cdc42 and Rac1 blocked upregulation of filopodia and augmentation of cell-surface area, respectively. Activation of these GTPases by netrin-1 was verified by detecting the amount of GTP-bound Cdc42 and Rac1 by western blot analysis. Interestingly, the authors showed that DCC was able to activate Cdc42 or Rac1 independently. A second study conducted by the group of Lamarche-Vane confirms the role of Cdc42 and Rac1 during neurite extension and provides evidence for the requirement for downregulation of RhoA and Rho-kinase [2xSee all References][2].Future studies will have to elucidate the molecular nature of the protein(s) coupling DCC to GTPases. Because DCC is considered to be a tumor-suppressor gene, alongside the fundamental interest generated by elucidating the signaling pathway stemming from its activation, these studies could well shed light on the mechanisms controlling tumor progression.

The autonomy of axons: no need for a cell body

Axon outgrowth is an intriguing process that is essential for the establishment of neuronal connections. Despite tremendous progress in the knowledge of the basic mechanisms ensuring axon elongation, many questions are unresolved. Among them, there has been much discussion about the capability of axons to perform protein synthesis. This is particularly important as the growth cones have to detect the pathway towards the final targets over long distances and must adapt their response as a function of the positional cues encountered, often far away from the cell body. The group of J.G. Flanagan [1xAxonal protein synthesis provides a mechanism for localized regulation at an intermediate target. Brittis, P.A. et al. Cell. 2002; 110: 223–235Abstract | Full Text | Full Text PDF | PubMed | Scopus (293)See all References[1] recently addressed this point in a very well-designed study that provides further evidence for axonal protein synthesis.The authors followed the growth of individual axons, separated from the cell body, after introduction of RNA constructs by viral vector or electroporation. These constructs were designed to synthesize green-fluorescent protein (GFP), allowing the visualization of translation within the tip of axons, or formed membrane-anchored alkaline phosphatase to determine whether synthesized proteins were exported to the cell surface. Strikingly, the authors demonstrated that axons synthesize proteins and address them to the cell surface in the absence of the cell body. Moreover, based on several studies demonstrating upregulation of various receptors for guidance molecules in axons after crossing the midline of the CNS (thereby modifying their responsiveness to this intermediate target), the authors analyzed the consequence of protein synthesis on axon behavior by introducing different RNA constructs into commissural axons in organotypic slices (open-book preparation). They uncovered a mechanism by which untranslated mRNA sequences are required for appropriate upregulation of protein synthesis in the distal part of growing axons at the midline floor plate.Overall, this study provides strong evidence that protein synthesis is not restricted to the cell body, that growing axons can upregulate protein synthesis once reaching an intermediate target and, finally, it highlights an RNA-based mechanism regulating this protein synthesis. The supremacy of the cell body previously considered as the ‘command center’ of the cell is now affected by the properties of axons that, far from having only the ‘simple’ function of protein transportation, appear as independent centers capable of protein synthesis and controlling that synthesis through sophisticated mechanisms that warrant elucidation.

Dissecting the role of Slit fragments

The midline of the central nervous system (CNS) is an extraordinary source for the discovery of new guidance molecules controlling axon pathways. Among the recent signals characterized, the Slits have been shown to be secreted glycoproteins with repulsive properties both in Drosophila and in mammals. These large molecules comprise four structural domains and can be proteolytically cleaved. This property is thought to underlie the differential bioactivity of the proteins. The N-terminal fragment of Slit has recently been considered as the functional domain evoking repulsion of olfactory bulb axons and subventricular zone neurons. Facing the lack of information concerning the functional activities of the different fragments, two groups designed elegant experiments dissecting the properties of each domain 1xRepellent signaling by Slit requires the leucine-rich repeats. Battye, R. et al. J. Neurosci. 2001; 21: 4290–4298PubMedSee all References, 2xDiversity of actions of slit2 proteolytic fragments in axon guidance. Nguyen Ba-Charvet, K.T. et al. J. Neurosci. 2001; 21: 4281–4289PubMedSee all References.In their study, Jacobs and colleagues 1xRepellent signaling by Slit requires the leucine-rich repeats. Battye, R. et al. J. Neurosci. 2001; 21: 4290–4298PubMedSee all References1 analyzed the effects of various mutations on the repellent signaling triggered by Slit. They found that point mutations in the leucine-rich repeat domain (LRR) reduced the repulsive property of Slit. Interestingly, they were able to generate Slit transgenes with specific deletions that were expressed in Slit mutants. Based on the ability of transgenes to restore a normal phenotype, the authors confirmed the importance of the LRR domain, which had to be intact for complete restoration. This approach also demonstrated that the LRR domain mediates binding to the major Slit receptor, Robo.A similar study has been conducted by Nguyen Ba-Charvet et al.2xDiversity of actions of slit2 proteolytic fragments in axon guidance. Nguyen Ba-Charvet, K.T. et al. J. Neurosci. 2001; 21: 4281–4289PubMedSee all References2, who focused on the vertebrate Slit2 that has been shown to be cleaved in vivo. Moreover, this study was motivated by experiments suggesting that one of the two generated fragments, the N-terminal fragment, possesses a branch-promoting activity on dorsal root ganglia (DRG) axons. In order to explore the role of the two fragments, the authors constructed mutant versions of Slit2 and examined the activities of truncated proteins on various axonal populations. Their results provide further evidence for a specific activity of the N-terminal fragment (containing the LRR motif) during both axon repulsion and branching of DRG neurons. The full-length uncleaved protein also exhibited the repulsive property but antagonized the branching effect, suggesting differences in the recruitment of the receptors mediating repulsion and branching. Strikingly, the authors report that the repulsion exerted by the full-length protein was not linked to growth cone collapse, providing exciting evidence for a dissociation of growth cone collapse and axon repulsion.The next challenge will be to understand the function of the C-terminal fragment of Slit that appeared inactive despite its epidermal growth factor-like (EGF) repeats and laminin-like globular domain that is largely involved in cell signaling and binding to numerous extracellular proteins.